Pinned
Launching $MTDV: Open-Source Privacy & Security Toolkit
This year I shipped 7 MIT-licensed tools under the $MTDV umbrella, focused on privacy-preserving tech, quantum security, and AI safety. All available at mtdv.orbiter.website
🔐 dotclaude - AI development security toolkit
→ github.com/ngmisl/dotc...

?
$MTDV
6h vol
$19K
3
8
13
Done. The panic button:
- Sits in the nav bar, styled as a red pill badge ("Panic") that goes solid red on hover
- On click, it immediately:
a. Clears textarea and all rendered results
b. Deletes model weights from IndexedDB, then deletes the entire database
c. Unregisters the service worker and clears al...
2
0
4
Concluding Klearu for my idea now (latest update)
Learnings:
- Fine-tuned dense works well, data classifications are passing (~50s on avg) - too slow for my specific usecase
- Sparse is still broken for Qwen, *however* it doesn't really make sense, sparse inference was designed for 7B+ models where MLP layers are 11...
0
0
5
Fantasma: A PWA where you paste text and a fine-tuned Qwen 0.8B model rewrites it to neutralize your stylometric fingerprint. Same meaning, no authorship signal. Runs entirely in-browser via Zig WASM. The tool that anonymizes your writing never sees a server.
1
0
3
Unfortunately I haven't been able to get better results (worse actually) using Klearu's sparse (patched) with Qwen3.5 0.8b - IMO the model is too small to make a diff
Full breakdown of my Klearu expedition:
The Good:
- Skills, when you got the token-space are more than good enough to go from zero to hero with small models
- Fine tuned with a small lite system prompt is even better, took me ~1hr to generate synthetic data and use unsloth to tune
- with vLLM both base+skills and fine

0
0
0
PSA: $MTDV Empire was upgraded to v4, which means new CA. Every tokenholder can claim their airdrop after initial 24h here: www.empirebuilder.world/empire/0xc50...
Thanks to @diviflyy for the support!
1
4
5
⛏️ Come mine Farbits with me!
Bitcoin reimagined for Farcaster. Mine every hour, random winner takes the block bounty. Proof of Cast on Base.

0
0
2
Full breakdown of my Klearu expedition:
The Good:
- Skills, when you got the token-space are more than good enough to go from zero to hero with small models
- Fine tuned with a small lite system prompt is even better, took me ~1hr to generate synthetic data and use unsloth to tune
- with vLLM both base+skills and fine...

1
2
8


